A method for on-line oil purification and performance regeneration of hydraulic oil by centrifugal field regulation

Abstract

This invention relates to the field of hydraulic oil purification technology, specifically a method for online hydraulic oil purification and performance regeneration based on centrifugal force field control. The method includes: online acquisition of real-time operating parameters of the hydraulic system and hydraulic oil state data; prediction of contaminant concentration distribution and deteriorated component content; dynamic calculation of the centrifugal force field intensity distribution curve using a state inversion algorithm; and control of the centrifugal device to form a gradient-changing directional centrifugal force field. Based on gradient centrifugation, the method removes solid particles, free water, and colloidal substances step-by-step. Simultaneously, it fine-tunes the local intensity of the centrifugal force field according to the deteriorated component content, selectively removing deteriorated oil molecular components while retaining intact base oil and effective additives. This method enables precise online grading and purification of hydraulic oil, adapts to the real-time state of the oil, improves impurity removal, simultaneously completes oil performance regeneration, and optimizes the quality of purification and regeneration treatment.

Description

Technical Field

[0001] This invention relates to the field of hydraulic oil purification technology, and in particular to an online method for purifying and regenerating hydraulic oil by means of centrifugal force field control. Background Technology

[0002] Current online hydraulic oil purification and performance regeneration processes generally employ conventional centrifugal separation technology. This relies on a fixed-parameter centrifugal force field to physically separate solid particles and free water from the hydraulic oil. However, this technology only collects basic hydraulic oil condition data without simultaneously integrating it with real-time operating parameters of the hydraulic system for comprehensive analysis. Consequently, it cannot quantitatively predict the concentration distribution of different contaminants or the content of performance-degrading components within the hydraulic oil, making it difficult to generate oil condition assessment results that accurately reflect actual operating conditions. Existing centrifugal purification equipment cannot dynamically calculate the centrifugal force field intensity distribution using algorithms; the centrifugal separation force remains constant without gradient changes. It cannot provide differentiated separation driving forces based on the physical characteristics of contaminants, and can only achieve uniform separation of conventional impurities.

[0003] Conventional centrifugal oil purification technology can only remove solid particles and free water, and its ability to handle colloidal substances and degraded oil components is insufficient. It cannot differentiate between degraded oil molecules and intact base oil and effective additives during the purification process, and it lacks the function of regenerating hydraulic oil performance. This invention addresses the problems of existing centrifugal force fields failing to dynamically match the real-time state of the oil, lacking gradient grading during separation, and being unable to selectively remove degraded components. It determines the centrifugal force field intensity distribution curve through real-time data fusion prediction and state inversion algorithms, achieving graded oil purification based on a directional gradient centrifugal force field, and simultaneously fine-tuning the local intensity of the centrifugal force field to selectively remove degraded components. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the existing technology and propose an online hydraulic oil purification and performance regeneration method controlled by centrifugal force field.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a method for online oil purification and performance regeneration of hydraulic oil controlled by centrifugal force field, comprising: Real-time operating parameters of the hydraulic system and real-time status data of the hydraulic oil are collected online. Based on the real-time operating parameters and real-time status data, the concentration distribution of different contaminants and the content of performance-degrading components in the hydraulic oil are predicted to obtain the oil condition assessment results. Based on the oil condition assessment results, the intensity distribution curve of the required centrifugal force field is dynamically calculated using a state inversion algorithm. The intensity distribution curve is used to guide the generation of gradient-changing centrifugal separation force. Based on the intensity distribution curve, the centrifuge device is controlled to generate a suitable directional centrifugal force field, which provides differentiated separation driving force according to the physical characteristics of the pollutants; Under the action of the directional centrifugal force field, the hydraulic oil is subjected to gradient centrifugal classification and purification treatment, and solid particles, free water and colloidal substances are removed step by step by adjusting the centrifugal force step by step. During the simultaneous process of gradient centrifugal grading for oil purification, the local intensity of the centrifugal force field is finely adjusted based on the content of performance-degraded components identified by the state inversion algorithm, in order to selectively remove degraded oil molecules while retaining intact base oil and effective additives.

[0006] As a further aspect of the present invention, based on the real-time operating condition parameters and the real-time status data, the concentration distribution of different contaminants and the content of performance-degrading components in the hydraulic oil are predicted, including: The real-time operating parameters include system pressure and oil temperature, and the real-time status data includes contamination level data, viscosity data, and acid value data. The real-time operating condition parameters and the real-time status data are input into a pre-trained pollutant association inference network, which is a computational model containing multiple nonlinear transformations. The pollutant association inference network maps the predicted concentration of solid particulate matter, the predicted content of emulsified water, and the predicted content of colloidal substances in different particle size ranges. Simultaneously, the viscosity data and acid value data from the real-time status data are input into the performance degradation assessment network, which is used to correlate changes in the chemical structure of the oil. The content of oxidized polymers, acidic substances, and additive loss products in the oil is calculated using the performance degradation assessment network, which is the content of the performance degradation components. The predicted concentration values ​​are combined with the content of the performance-degrading components to form a complete oil condition assessment result.

[0007] As a further aspect of the present invention, based on the oil condition assessment results, the intensity distribution curve of the required centrifugal force field is dynamically calculated using a state inversion algorithm, including: The predicted values ​​of pollutant concentrations and the content of performance-degrading components in the oil condition assessment results are used as the input vectors for the condition inversion algorithm. Establish a set of inversion equations describing the relationship between centrifugal force field intensity, interaction time, and removal efficiency of specific pollutants; Substitute the input vector into the inversion equation set and solve in reverse the process to obtain the centrifugal force field intensity required to achieve the target removal efficiency as a function of time or spatial position; The obtained function is smoothed and piecewise fitted to generate the intensity distribution curve that changes continuously within a preset processing period. The intensity distribution curve clarifies the intensity value that the centrifugal force field should reach in different processing stages.

[0008] As a further aspect of the present invention, based on the intensity distribution curve, controlling the centrifugal device to generate a suitable directional centrifugal force field includes: The intensity distribution curve is converted into a sequence of spindle speed control commands for the centrifugal device. According to the spindle speed control command sequence, the speed of the drive motor is adjusted so that the centrifugal drum produces a speed change that matches the intensity distribution curve; Inside the centrifugal drum, based on the characteristics of the difference in centrifugal force at different radii, and combined with the density and particle size information of the pollutants, the optimal separation zone for different pollutants is planned. By adjusting the spatial position and flow field distribution of the oil inlet, the hydraulic oil to be treated is guided to the corresponding optimal separation area, thereby forming a directional centrifugal force field inside the centrifugal drum that matches the contamination characteristics.

[0009] As a further aspect of the present invention, under the action of the directional centrifugal force field, the hydraulic oil undergoes gradient centrifugal classification and purification treatment, including: In the initial stage of gradient centrifugal classification oil purification, a relatively low initial centrifugal force field intensity is used to allow the densest and largest solid particles in the hydraulic oil to settle and separate first in the directional centrifugal force field. After the initial stage is completed, the centrifugal force field intensity is increased to the intermediate intensity range according to the intensity distribution curve; Under the aforementioned medium strength range, fine particles with low density, demulsified free water, and some macromolecular colloidal substances are removed from the hydraulic oil; In the final stage of gradient centrifugal grading oil purification, the intensity of the centrifugal force field is adjusted to the highest intensity within a preset range according to the intensity distribution curve. At the highest intensity, residual fine particles, emulsified water, and soluble colloids are further removed to complete the graded oil purification process.

[0010] As a further aspect of the present invention, during the synchronous process of gradient centrifugal grading and oil purification, the local intensity of the centrifugal force field is finely adjusted based on the content of performance-degrading components identified by the state inversion algorithm, including: In the final stage of the gradient centrifugal classification oil purification process, a simultaneous separation procedure for performance-degraded components is initiated. Based on the content of performance-deteriorating components in the oil condition assessment results, calculate the additional centrifugal force increment required for selective separation; Based on the maximum intensity specified by the intensity distribution curve, the additional centrifugal force increment is superimposed to form a reinforced separation force field targeting the performance-deteriorating component; Under the enhanced separation force field, oxidized and polymerized macromolecules, some acidic products, and ineffective additives are enriched and separated during centrifugation, while intact base oil and effective additives are retained in the purified oil.

[0011] As a further aspect of the present invention, it also includes: Real-time monitoring of the cleanliness and key performance indicators of hydraulic oil after gradient centrifugal grading purification, generating feedback data of the purification process; The feedback data from the oil purification process and the oil condition assessment results are input into the collaborative control logic model, which is used to optimize the centrifugal force field control strategy. The intensity distribution curve and the operating parameters of the gradient centrifugal grading oil purification process are adjusted in a closed loop using the aforementioned collaborative control logic model. The process is iteratively executed until the cleanliness and key performance indicators of the hydraulic oil reach the preset standards, thus completing the online oil purification and performance regeneration treatment of the hydraulic oil. The real-time monitoring of the cleanliness and key performance indicators of the hydraulic oil after gradient centrifugal grading purification includes: An online particle counter and a moisture sensor are installed in the clean oil outlet pipeline of the centrifuge to continuously acquire the number of particles and the water content of the treated hydraulic oil as the cleanliness index. A portion of oil sample is drawn out from the upstream branch of the clean oil outlet pipeline and flows through an online viscometer and acid value sensor to measure the kinematic viscosity and total acid value of the treated hydraulic oil, which are used as the key performance indicators. The data streams of particle number, water content, kinematic viscosity and total acid value are timestamped and packaged to generate the oil purification process feedback data.

[0012] As a further aspect of the present invention, the feedback data from the oil purification process and the oil condition assessment results are input together into the collaborative control logic model, including: An error assessment mechanism is established that uses the feedback data from the oil purification process as real-time observations and the oil condition assessment results as prediction benchmarks. By comparing the real-time observed values ​​with the predicted baseline using the collaborative control logic model, the deviation between the current oil purification effect and the expected target is calculated. Based on the calculated deviation, determine whether the deviation is due to insufficient centrifugal force field strength, improper grading sequence, or inaccurate identification of performance-degrading components. Based on the judgment of the source of deviation, the collaborative control logic model generates a set of control instructions including the adjustment amount of centrifugal force field intensity, the adjustment amount of graded stage duration, or the correction amount of performance degradation component identification parameters.

[0013] As a further aspect of the present invention, the intensity distribution curve and the operating parameters of the gradient centrifugal grading oil purification process are adjusted in a closed loop through the aforementioned collaborative control logic model, including: The centrifugal force field intensity adjustment amount in the control instruction set is fed back to the input parameters of the state inversion algorithm, triggering the state inversion algorithm to recalculate the new intensity distribution curve; The duration adjustment of the grading stage is directly applied to the timing controller of the gradient centrifugal grading oil purification process to adjust the duration of each stage of the process. The correction amount of the performance degradation component identification parameter is input into the internal weight matrix of the inference calculation model to fine-tune the performance degradation assessment network online, so as to improve the prediction accuracy of the content of performance degradation components. Based on the newly generated intensity distribution curve, the adjusted treatment duration, and the fine-tuned performance degradation assessment network, a new round of online hydraulic oil purification and performance regeneration treatment cycle is initiated.

[0014] As a further aspect of the present invention, the steps for constructing the collaborative control logic model include: Construct a closed-loop control network architecture with the feedback data of the oil purification process as the observation input, the oil condition assessment results as the reference input, and the control commands as the output; In the closed-loop control network architecture, a state comparator is designed to compare the real-time observed values ​​of cleanliness indicators and key performance indicators in the oil purification process feedback data with the corresponding predicted values ​​of pollutant concentration and performance degradation component content in the oil condition assessment results, and calculate the deviation vector between each predicted value and the real-time observed value. Based on the aforementioned deviation vector, the weighted influence coefficients of each deviation on the oil purification effect are calculated using a fuzzy inference engine. Using a multi-objective decision-making algorithm, the weighted influence coefficients, preset energy consumption constraints, and processing efficiency constraints are comprehensively calculated to generate a candidate instruction set that includes the centrifugal force field intensity adjustment, the graded stage duration adjustment, and the performance degradation component identification parameter correction. The strategy evaluation module simulates and evaluates the execution effect of each control instruction in the candidate instruction set, selects the optimal control instruction combination that makes the predicted processing result of the next processing cycle approach the preset standard the fastest, and outputs the optimal control instruction combination as the control instruction set of the collaborative control logic model.

[0015] Compared with the prior art, the advantages and positive effects of the present invention are as follows: Real-time acquisition of hydraulic system operating parameters and hydraulic oil status data enables quantitative prediction of the concentration distribution of different contaminants and the content of performance-degrading components in the hydraulic oil. This results in an oil condition assessment that closely matches the actual operating conditions. A state inversion algorithm dynamically calculates a centrifugal force field intensity distribution curve that matches the oil condition. This curve directly guides the generation of gradient-varying centrifugal separation forces, controlling the centrifugal device to form a directional centrifugal force field that matches the oil condition. This directional centrifugal force field provides differentiated separation driving forces based on the physical characteristics of the contaminants. Combined with gradient centrifugal grading and oil purification, it can sequentially remove solid particles, free water, and colloidal substances through progressively adjusted centrifugal forces. The application of centrifugal force corresponds to the type and distribution of contaminants, avoiding the insufficient separation caused by a single fixed centrifugal force field. This ensures that contaminants with different physical characteristics can be effectively separated under appropriate centrifugal force, improving the removal rate of various impurities in the hydraulic oil.

[0016] During the simultaneous gradient centrifugal grading and purification process, the local intensity of the centrifugal force field can be precisely fine-tuned based on the content of performance-degraded components identified by the state inversion algorithm. By adjusting the parameters of the local centrifugal force field, the directional removal of degraded oil molecules is achieved. During the separation of degraded components, the intact base oil and effective additives are not separated, thus fully preserving the effective components in the hydraulic oil. This treatment method can simultaneously restore the performance of hydraulic oil during physical purification, breaking through the limitations of conventional centrifugal purification which only achieves physical impurity separation. It restores oil performance without the need for additional regeneration reagents, reducing the loss of effective components in the hydraulic oil. By simultaneously regenerating oil performance during purification, it restores degraded hydraulic oil to its original performance, achieving integrated online purification and performance regeneration of hydraulic oil. Attached Figure Description

[0017] Figure 1 This is a flowchart of an online hydraulic oil purification and performance regeneration method controlled by centrifugal force field according to the present invention; Figure 2 A flowchart for controlling the centrifuge device to generate a directional centrifugal force field based on the intensity distribution curve; Figure 3 A curve showing the improvement in the closed-loop control and iterative processing effect; Figure 4 The diagram shows the centrifugal force field intensity distribution curve and control command diagram; Figure 5 This is a deviation pattern recognition diagram for a hydraulic oil gradient centrifugal purification system. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0019] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0020] See Figure 1 This invention provides a method for online oil purification and performance regeneration of hydraulic oil controlled by centrifugal force field, the specific method including: Real-time operating parameters of the hydraulic system and real-time state data of the hydraulic oil are collected online. Based on this real-time data, the concentration distribution of different contaminants and the content of performance-degrading components in the hydraulic oil are predicted, resulting in an oil condition assessment. According to this oil condition assessment, the intensity distribution curve of the required centrifugal force field is dynamically calculated using a state inversion algorithm. This curve guides the generation of gradient-varying centrifugal separation forces. Based on the calculated intensity distribution curve, the centrifugal device is controlled to generate an appropriate directional centrifugal force field, which provides differentiated separation driving forces based on the physical characteristics of the contaminants. Under the action of this directional centrifugal force field, the hydraulic oil undergoes gradient centrifugal grading and purification treatment, removing solid particles, free water, and colloidal substances step by step through progressively adjusted centrifugal forces. Simultaneously during this gradient centrifugal grading and purification process, the local intensity of the centrifugal force field is fine-tuned based on the content of performance-degrading components identified by the state inversion algorithm to selectively remove degraded oil molecules while retaining intact base oil and effective additives.

[0021] In one embodiment of the present invention, real-time operating parameters include system pressure and oil temperature, and real-time status data includes contamination level data, viscosity data, and acid value data. These real-time operating parameters and real-time status data are input into a pre-trained contaminant association inference network, which is a computational model incorporating multi-layer nonlinear transformations. Through this contaminant association inference network, predicted concentrations of solid particles, predicted emulsion water content, and predicted colloidal substance content are mapped across different particle size ranges. Simultaneously, viscosity data and acid value data from the real-time status data are input into a performance degradation assessment network, which is used to correlate changes in the chemical structure of the oil. Through this performance degradation assessment network, the contents of oxidized polymers, acidic substances, and additive loss products in the oil are calculated, i.e., the contents of performance degradation components. Integrating the obtained concentration predictions and performance degradation component contents constitutes a complete oil condition assessment result.

[0022] In practical implementation, real-time operating parameters include system pressure and oil temperature, and real-time status data includes contamination level data, viscosity data, and acid value data. These real-time operating parameters and real-time status data are input into a pre-trained contaminant association inference network. The contaminant association inference network is a computational model containing multiple layers of nonlinear transformations, and the mapping relationship of the contaminant association inference network is expressed by the following formula:

[0023] in: The input vector consists of system pressure, oil temperature, contamination level data, viscosity data, and acid value data. This represents the output vector, which contains the predicted concentration of solid particles, the predicted content of emulsion water, and the predicted content of colloidal substances in different particle size ranges. and The weight matrix represents the pollutant association inference network; and The bias vector represents the pollutant association inference network; This represents the activation function of the pollutant association inference network. Through this network, predicted concentrations of solid particulate matter, emulsion water content, and colloidal matter content across different particle size ranges are mapped. Simultaneously, viscosity and acid value data from real-time status data are input into the performance degradation assessment network. This network correlates changes in the oil's chemical structure and calculates the content of oxidized polymers, acidic substances, and additive loss products in the oil—that is, the content of performance degradation components. Integrating the predicted concentrations and the content of performance degradation components yields a complete oil condition assessment result.

[0024] In some embodiments, the training of the contaminant association inference network relies on a historical hydraulic oil condition dataset, which includes system pressure, oil temperature, contamination level, viscosity, and acid value data, as well as corresponding laboratory-detected contaminant concentrations. The training process updates network parameters by minimizing the error between the predicted output and the true value. It can be understood that the multi-layer nonlinear transformation structure of the contaminant association inference network allows the model to capture the complex nonlinear relationship between input features and contaminant concentrations, thereby improving prediction accuracy. In a specific implementation, the performance degradation assessment network adopts a similar multi-layer neural network architecture, but the input is limited to viscosity and acid value data, and the output consists of the content values ​​of three performance degradation components: oxidized polymer content, acidic substance content, and additive loss products content. The performance degradation assessment network is trained based on the chemical analysis results of the oil sample, and the internal weights of the network are adjusted through supervised learning. Optionally, the contaminant association inference network and the performance degradation assessment network can be deployed on embedded industrial controllers or edge computing devices to achieve real-time data input and prediction calculations to support online oil purification processes. It is understandable that the integration operation of the oil condition assessment results is to combine the predicted values ​​of solid particulate matter concentration, emulsion water content, and colloidal substance content output by the pollutant association inference network with the content of oxidized polymers, acidic substances, and additive loss products output by the performance degradation assessment network into a comprehensive data vector. This comprehensive data vector is directly passed to the subsequent condition inversion algorithm as the oil condition assessment result.

[0025] In one embodiment of the present invention, the predicted contaminant concentration and the content of performance-degrading components in the oil condition assessment results are used as the input vectors for the condition inversion algorithm. An inversion equation set describing the relationship between centrifugal force field intensity, treatment time, and the removal efficiency of a specific contaminant is established. The input vector is substituted into the inversion equation set to solve inversely for the centrifugal force field intensity required to achieve the target removal efficiency as a function of time or spatial location. The solved function is smoothed and piecewise fitted to generate a continuously varying intensity distribution curve within a preset treatment cycle. This intensity distribution curve clearly defines the intensity values ​​that the centrifugal force field should achieve at different treatment stages. (See also...) Figure 2 The intensity distribution curve is converted into a sequence of spindle speed control commands for the centrifuge device. Based on this sequence, the speed of the drive motor is adjusted to cause the centrifugal drum to exhibit speed variations that match the intensity distribution curve. Inside the centrifugal drum, based on the characteristics of centrifugal force differences at different radii and combined with information on pollutant density and particle size, optimal separation zones for different pollutants are planned. By adjusting the spatial position and flow field distribution of the oil inlet, the hydraulic oil to be treated is guided to the corresponding optimal separation zone, thereby forming a directional centrifugal force field inside the centrifugal drum that matches the characteristics of the pollutants.

[0026] In practical implementation, the predicted contaminant concentration and the content of performance-degrading components from the oil condition assessment results are used as the input vectors for the state inversion algorithm. A set of inversion equations describing the relationship between centrifugal force field intensity, application time, and the removal efficiency of specific contaminants is established. The core solution relationship of the state inversion algorithm can be expressed by the following formula:

[0027] in: Indicates the position in time or processing sequence. The required centrifugal force field strength at that location; This represents the pollutant prediction vector, which includes the predicted concentration of solid particulate matter in different particle size ranges, the predicted content of emulsified water, and the predicted content of colloidal substances. This represents a vector indicating the content of components that degrade performance, including the content of oxidized polymers, acidic substances, and additive loss products. This represents the set of internal parameters of the state inversion algorithm, which is obtained through training with historical data. This represents the nonlinear inversion operator defined in the state inversion algorithm, which encapsulates the mathematical relationships of the inversion equations. Substituting the input vector into the inversion equations, the algorithm solves inversely for the centrifugal force field intensity required to achieve the target removal efficiency as a function of time or spatial location. The solved function is smoothed and piecewise fitted to generate a continuously varying intensity distribution curve within a preset processing period. This curve clearly defines the intensity values ​​that the centrifugal force field should achieve at different processing stages. The intensity distribution curve is then converted into a sequence of spindle speed control commands for the centrifuge device. This conversion process maps the target centrifugal force intensity value to the corresponding target spindle speed value based on the physical relationship between centrifugal force and rotational speed, and arranges these commands into a time sequence.

[0028] In some embodiments, the inversion equation set is established based on Stokes' law of centrifugal sedimentation and the pollutant removal kinetic model, and the internal parameter set of the state inversion algorithm. The optimization is achieved by processing a large number of paired samples of known contamination states and optimal centrifugal force field parameters. It can be understood that the nonlinear inversion operator of the state inversion algorithm... It can handle the coupling relationship between multi-component pollutants and performance-degrading components in the input vector, thereby calculating an intensity distribution curve that can simultaneously optimize the removal effect of multiple pollutants. In specific implementation, the speed of the drive motor is adjusted according to the spindle speed control command sequence, so that the centrifugal drum produces a speed change that matches the intensity distribution curve. The control of the drive motor is realized through a closed-loop variable frequency speed regulation system. The system reads the actual spindle speed in real time and compares it with the target speed in the control command sequence. The motor output is adjusted through a proportional-integral-derivative algorithm to ensure that the speed response accurately follows the intensity distribution curve. Inside the centrifugal drum, based on the characteristics of the difference in centrifugal force at different radii, combined with the density and particle size information of the pollutants, the optimal separation zone for different pollutants is planned. The planning process is completed by calculating the settling velocity and residence time of particles with specific densities and sizes at different radial positions inside the drum. It can be understood that the optimal separation zone is a spatial range associated with the physical properties of the pollutants. Applying centrifugal force within this range, as specified by the intensity distribution curve, can achieve the highest separation efficiency for the target pollutants. Optionally, by adjusting the spatial position and flow field distribution of the oil inlet, the hydraulic oil to be treated can be guided to the corresponding optimal separation area. The oil inlet is usually designed to be movable radially along the drum or has a multi-channel adjustable structure. The flow field distribution is adjusted by the guide vanes or flow rate controller in the oil inlet pipeline, thereby forming a directional centrifugal force field inside the centrifugal drum that matches the contamination characteristics.

[0029] In one embodiment of the present invention, in the initial stage of the gradient centrifugal classification and purification process, a relatively low initial centrifugal force field intensity is used to allow the densest and largest solid particles in the hydraulic oil to settle and separate first in the directional centrifugal force field. After the initial stage is completed, the centrifugal force field intensity is increased to a medium intensity range according to the intensity distribution curve. At the medium intensity range, fine particles with low density, demulsified free water, and some large molecular colloidal substances in the hydraulic oil are removed. In the final stage of the gradient centrifugal classification and purification process, the centrifugal force field intensity is adjusted to the highest intensity within a preset range according to the intensity distribution curve. At the highest intensity, residual fine particles, emulsified water, and dissolved colloids are further removed, completing the classification and purification process. In the final stage of the gradient centrifugal classification and purification process, a synchronous separation procedure for performance-deteriorated components is initiated. Based on the content of performance-deteriorated components in the oil condition assessment results, the additional centrifugal force increment required for selective separation is calculated. Based on the highest intensity specified by the intensity distribution curve, this additional centrifugal force increment is superimposed to form a strengthened separation force field for performance-deteriorated components. Under the action of the enhanced separation force field, the oxidized and polymerized macromolecules, some acidic products and ineffective additives are enriched and separated during the centrifugation process, while the intact base oil and effective additives are retained in the purified oil.

[0030] In practice, the initial stage of the gradient centrifugal classification oil purification process uses a relatively low initial centrifugal force field intensity. The initial centrifugal force field intensity is set based on the intensity distribution curve's requirements for separating large-sized solid particles, ensuring that the densest and largest solid particles in the hydraulic oil settle and separate first in the directional centrifugal force field. After the initial stage, the centrifugal force field intensity is increased to a medium intensity range based on the intensity distribution curve. The lower limit of the medium intensity range is higher than the initial centrifugal force field intensity, and the upper limit is lower than the highest intensity within the preset range. At the medium intensity range, fine particles with low density, demulsified free water, and some large-molecule colloidal substances are removed from the hydraulic oil. In the final stage of the gradient centrifugal classification oil purification process, the centrifugal force field intensity is adjusted to the highest intensity within the preset range based on the intensity distribution curve. At the highest intensity, residual fine particles, emulsified water, and dissolved colloids are further removed, completing the classification oil purification process. See Table 1 for parameter examples for different stages.

[0031] Table 1: Example Table of Parameters for Gradient Centrifugation and Classification Oil Purification Stages ; In some embodiments, during the final stage of gradient centrifugal classification and purification, a simultaneous separation procedure for performance-deteriorating components is initiated. Based on the content of performance-deteriorating components in the oil condition assessment results, the additional centrifugal force increment required for selective separation is calculated. The calculation of the additional centrifugal force increment is expressed by the following formula:

[0032] in: This indicates the additional centrifugal force increment required for the selective separation of components with deteriorating performance; This represents a vector indicating the content of components that degrade performance, including the content of oxidized polymers, acidic substances, and additive loss products. This represents the set of internal parameters of the performance degradation assessment network; This represents the functional mapping relationship for calculating the incremental increase in centrifugal force based on the content of performance-deteriorating components. Based on the maximum intensity specified by the intensity distribution curve, the incremental increase in centrifugal force is superimposed to form a reinforced separation force field targeting the performance-deteriorating components. The intensity of this reinforced separation force field is the sum of the maximum intensity and the incremental increase in centrifugal force. Under the action of this reinforced separation force field, oxidized and polymerized macromolecules, some acidic products, and ineffective additives are enriched and separated during centrifugation, while intact base oil and effective additives are retained in the purified oil.

[0033] It is understandable that gradient centrifugal grading for oil purification achieves the step-by-step separation of contaminants through progressively increasing centrifugal force field intensity. The initial stage, with its lower centrifugal force field intensity, aims to efficiently remove large, high-density particles to avoid interference with subsequent separation stages. Optionally, the duration of each treatment stage is not fixed but is dynamically adjusted based on a time function defined in the intensity distribution curve or through real-time oil quality sensor feedback. In specific implementations, the simultaneous separation procedure of performance-degraded components overlaps temporally with the final high-intensity purification treatment, but the applied centrifugal force field intensities differ. The enhanced separation force field is an incremental field additionally applied to the base centrifugal force field in the final stage. This can be understood as an additional centrifugal force increment. Size and performance degradation component content vector The components are positively correlated; the higher the content of the performance-degrading component, the greater the calculated increase in additional centrifugal force. The larger the value, the stronger the applied enhanced separation force field. Optionally, the function mapping relationship... The specific form can be obtained by training on successful separation data of historical performance-deteriorated oil samples using machine learning methods to ensure the accuracy of the calculation of the additional centrifugal force increment.

[0034] See Figure 3 This is a curve illustrating the improvement in the effect of closed-loop control iteration, used to demonstrate the impact of the number of closed-loop control iterations on the overall state of hydraulic oil in an online oil purification and performance regeneration system. Both curves show a continuous upward trend followed by gradual convergence, proving the effectiveness of the closed-loop control model. The more iterations, the closer the cleanliness and performance recovery level of the oil are to the preset qualified standards. In the initial iteration, when the system starts processing for the first time, the basic centrifugal force field can only complete the preliminary classification and purification, with limited effects on the separation of deteriorated components and performance repair. In the mid-term iteration, the collaborative control logic model begins to dynamically adjust the intensity of the centrifugal force field and the classification time based on feedback data, rapidly improving the efficiency of contaminant removal. In the later iteration, as convergence approaches, the enhanced separation force field precisely acts on the deteriorated components, and the oil indicators basically reach the preset qualified standards, completing the closed-loop control.

[0035] In one embodiment of the present invention, the cleanliness and key performance indicators of the hydraulic oil after gradient centrifugal grading purification are monitored in real time to generate purification process feedback data. An online particle counter and a moisture sensor are installed in the clean oil outlet pipeline of the centrifuge to continuously acquire the particle count and water content of the treated hydraulic oil as cleanliness indicators. A portion of the oil sample is drawn from an upstream branch of the clean oil outlet pipeline and flows through an online viscometer and acid value sensor to measure the kinematic viscosity and total acid value of the treated hydraulic oil as key performance indicators. The data streams of particle count, water content, kinematic viscosity, and total acid value are timestamped and packaged to generate purification process feedback data. The purification process feedback data and the oil condition assessment results are input into a collaborative control logic model, which is used to optimize the centrifugal force field control strategy. The construction steps of the collaborative control logic model include: constructing a closed-loop control network architecture with the purification process feedback data as the observation input, the oil condition assessment results as the reference input, and the control commands as the output. In the closed-loop control network architecture, a state comparator is designed to compare the real-time observed values ​​of cleanliness and key performance indicators in the oil purification process feedback data with the corresponding predicted values ​​of contaminant concentration and performance degradation component content in the oil condition assessment results, calculating the deviation vector between each predicted value and the real-time observed value. Based on the deviation vector, a fuzzy inference engine calculates the weighted influence coefficients of each deviation on the oil purification effect. Using a multi-objective decision algorithm, the weighted influence coefficients, preset energy consumption constraints, and processing efficiency constraints are comprehensively calculated to generate a candidate instruction set including centrifugal force field intensity adjustment, grading stage duration adjustment, and performance degradation component identification parameter correction. The strategy evaluation module simulates and evaluates the execution effect of each control instruction in the candidate instruction set, selecting the optimal control instruction combination that makes the predicted processing result of the next processing cycle approach the preset standard the fastest. This optimal control instruction combination is output as the control instruction set of the collaborative control logic model.

[0036] In practical implementation, the cleanliness and key performance indicators of the hydraulic oil after gradient centrifugal grading purification are monitored in real time to generate purification process feedback data. An online particle counter and moisture sensor are installed in the clean oil outlet pipeline of the centrifuge to continuously acquire the particle count and water content of the treated hydraulic oil as cleanliness indicators. A portion of the oil sample is drawn from an upstream branch of the clean oil outlet pipeline and flows through an online viscometer and acid value sensor to measure the kinematic viscosity and total acid value of the treated hydraulic oil as key performance indicators. The data streams of particle count, water content, kinematic viscosity, and total acid value are timestamped and packaged to generate purification process feedback data. An example of the real-time monitored indicators and sensor deployment is shown in Table 2.

[0037] Table 2: Examples of Real-Time Monitoring Indicators and Sensor Configurations ; The feedback data from the oil purification process and the oil condition assessment results are input into the collaborative control logic model. This model optimizes the centrifugal force field control strategy. The construction steps of the collaborative control logic model include: building a closed-loop control network architecture with the oil purification process feedback data as the observation input, the oil condition assessment results as the reference input, and control commands as the output. In the closed-loop control network architecture, a state comparator is designed to compare the real-time observed values ​​of cleanliness indicators and key performance indicators from the oil purification process feedback data with the corresponding predicted values ​​of contaminant concentration and performance degradation component content from the oil condition assessment results, calculating the deviation vector between each predicted value and the real-time observed value. Based on the deviation vector, a fuzzy inference engine calculates the weighted influence coefficients of each deviation on the oil purification effect. Using a multi-objective decision-making algorithm, the weighted influence coefficients, preset energy consumption constraints, and treatment efficiency constraints are comprehensively calculated to generate a candidate command set that includes centrifugal force field intensity adjustment, grading stage duration adjustment, and performance degradation component identification parameter correction. The calculation relationship of the multi-objective decision-making algorithm can be expressed by the following formula:

[0038] in: This represents the set of candidate instructions generated; This represents the vector of weighted influence coefficients for each deviation calculated by the fuzzy inference engine. This represents the deviation vector calculated by the state comparator; This represents a pre-defined set of constraints, including energy consumption constraints and processing efficiency constraints. This represents the decision function of the multi-objective decision-making algorithm. The strategy evaluation module simulates and evaluates the execution effects of each control instruction in the candidate instruction set, selecting the optimal control instruction combination that makes the predicted processing result of the next processing loop approach the preset standard most quickly. This optimal control instruction combination is then output as the control instruction set of the collaborative control logic model.

[0039] In some embodiments, the state comparator calculates the deviation vector by subtracting the corresponding predicted value from each real-time observation, resulting in a numerical vector containing deviations in particle number, water content, kinematic viscosity, and total acid number. It can be understood that the rule base of the fuzzy inference engine is built based on expert knowledge or historical data training, used to map specific numerical deviations into weighted influence coefficients reflecting their impact on the final oil quality. In specific implementations, the multi-objective decision-making algorithm can be a non-dominated sorting genetic algorithm or a Pareto-optimal solution set search method, with the decision function... The goal is to satisfy the set of constraints. Under the premise of maximizing processing efficiency while minimizing energy consumption and time, the system seeks control commands that maximize processing efficiency while minimizing energy consumption and time. Optionally, the strategy evaluation module uses a simplified oil purification process simulation model to predict the oil state after the execution of different candidate commands. The simulation model is based on centrifugal separation kinetics and empirical formulas for contaminant removal. It can be understood that the closed-loop control network architecture, state comparator, fuzzy inference engine, multi-objective decision algorithm, and strategy evaluation module of the collaborative control logic model can be integrated into an industrial control computer or programmable logic controller to achieve online generation and output of control commands.

[0040] See Figure 4 This is a graph showing the distribution curve of centrifugal force field intensity and the control command diagram, fully illustrating the coordinated changes in key process parameters within a 120-minute oil purification cycle. The target centrifugal force exhibits a smoothly rising S-shaped curve, gradually increasing from an initial 5000g to a final 14800g, perfectly matching the "gradient increase" design logic in the patent, providing precise driving force for staged oil purification. The actual centrifugal force fluctuates slightly around the target curve, rapidly following the target value in the early stage (0-40min), exhibiting periodic oscillations in the middle stage (40-100min), and undergoing slight adjustments around the target value in the later stage (after 100min), verifying the tracking performance of the closed-loop control system. The spindle speed is strongly positively correlated with the centrifugal force field intensity, continuously rising from an initial 1000rpm to a peak of 2900rpm before stabilizing. The adjustment amount for the grading duration remains near 0 throughout, indicating that the current process parameters are close to optimal, requiring no significant adjustment to the stage duration, thus verifying the rationality of the initial process design.

[0041] In one embodiment of the present invention, an error assessment mechanism is established, using feedback data from the oil purification process as real-time observations and oil condition assessment results as prediction benchmarks. A collaborative control logic model compares the real-time observations with the prediction benchmarks to calculate the deviation between the current oil purification effect and the expected target. Based on the calculated deviation, the source of the deviation is determined to be insufficient centrifugal force field strength, improper grading sequence, or inaccurate identification of performance-degrading components. Based on the determination of the source of the deviation, the collaborative control logic model generates a set of control instructions including adjustments to centrifugal force field strength, grading stage duration, or corrections to performance-degrading component identification parameters. The centrifugal force field strength adjustment in the control instruction set is fed back to the input parameters of the state inversion algorithm, triggering the algorithm to recalculate a new intensity distribution curve. The grading stage duration adjustment is directly applied to the timing controller of the gradient centrifugal grading oil purification process to adjust the duration of each stage. The correction of performance-degrading component identification parameters is input into the internal weight matrix of the inference calculation model to fine-tune the performance degradation assessment network online, thereby improving the prediction accuracy of the content of performance-degrading components. Based on the newly generated intensity distribution curve, the adjusted treatment duration, and the fine-tuned performance degradation assessment network, a new round of online hydraulic oil purification and performance regeneration treatment cycle is initiated.

[0042] In practical implementation, an error assessment mechanism is established that uses the feedback data of the oil purification process as the real-time observation value and the oil condition assessment result as the prediction benchmark. The collaborative control logic model compares the real-time observation value with the prediction benchmark to calculate the deviation between the current oil purification effect and the expected target. After each treatment cycle, the collaborative control logic model reads the real-time observation values ​​of particle number, water content, kinematic viscosity and total acid value from the feedback data of the oil purification process, and reads the corresponding predicted values ​​of pollutant concentration and performance degradation component content from the oil condition assessment result. Then, the absolute difference or relative difference between the real-time observation value and the predicted value of each indicator is calculated to form a deviation vector containing multiple deviation components. Based on the calculated deviations, the source of the deviation is determined whether it is due to insufficient centrifugal force field strength, improper grading sequence, or inaccurate identification of performance-deteriorating components. The determination of the deviation source is based on the analysis of the deviation vector pattern. For example, if the real-time observed values ​​of particle number and water content are significantly higher than the predicted values, while the deviations of kinematic viscosity and total acid value are within acceptable ranges, the source of the deviation may be judged as insufficient removal of solid particles and water, pointing to insufficient centrifugal force field strength or improper grading sequence. If the real-time observed values ​​of kinematic viscosity and total acid value are significantly worse than the predicted values, while the cleanliness index deviation is small, the source of the deviation may be judged as poor separation effect of chemically deteriorating components in the oil, pointing to inaccurate identification of performance-deteriorating components leading to improper selective separation force field settings. Based on the determination of the deviation source, the collaborative control logic model generates a control instruction set including adjustments to centrifugal force field strength, grading stage duration, or performance-deteriorating component identification parameters. The generation of the control instruction set can be achieved through a decision function.

[0043] in: This represents the generated set of control instructions vectors; This represents the deviation vector calculated by the collaborative control logic model; This represents the deviation source identification vector obtained based on the deviation vector pattern judgment; This represents the mapping function from deviation analysis to the generation of specific adjustment instructions. The centrifugal force field intensity adjustment amount in the control instruction set is fed back to the input parameters of the state inversion algorithm, triggering the algorithm to recalculate a new intensity distribution curve. The feedback process is achieved by correcting the input parameters of the state inversion algorithm. For example, when the deviation is determined to be due to insufficient centrifugal force field intensity, the centrifugal force field intensity adjustment amount in the control instruction set will act as a correction coefficient on the target removal efficiency parameter in the state inversion algorithm, thereby driving the algorithm to calculate a new intensity distribution curve with higher force field intensity. The stage duration adjustment amount is directly applied to the timing controller of the gradient centrifugal staged oil purification process, adjusting the duration of each stage of treatment. Based on the received stage duration adjustment amount, the timing controller extends or shortens the control timing corresponding to the initial, intermediate, or final stages, thereby changing the centrifugal separation time of each pollutant component. The correction values ​​for the performance degradation component identification parameters are input into the internal weight matrix of the performance degradation assessment network to fine-tune the network online, thereby improving the prediction accuracy of the performance degradation component content. These correction values ​​are typically expressed as gradients or incremental matrices and are updated before the next forward computation cycle of the performance degradation assessment network. Based on the newly generated intensity distribution curve, the adjusted processing duration, and the fine-tuned performance degradation assessment network, a new cycle of online hydraulic oil purification and performance regeneration is initiated.

[0044] In some embodiments, the deviation source identification vector It is a multi-bit binary encoded vector, where each bit represents whether a possible source of bias is activated. This can be understood as a mapping function. Internally, a lookup table based on rules or empirical parameters is encapsulated. This lookup table defines the specific adjustment command values ​​corresponding to different combinations of deviation vector patterns and deviation source identification vectors. In implementation, after the control command set is generated, it undergoes formatting processing, transforming it into standardized control commands or parameter packages that can be directly recognized and executed by the state inversion algorithm, timing controller, and performance degradation assessment network. Optionally, the numerical values ​​of the centrifugal force field intensity adjustment, the grading stage duration adjustment, and the performance degradation component identification parameter correction are related to the deviation vector. The magnitudes of the corresponding deviation components are proportional; the larger the deviation, the larger the absolute value of the generated adjustment. It can be understood that inputting the correction amount of the performance degradation component identification parameters into the internal weight matrix of the performance degradation assessment network for online fine-tuning is a model parameter adaptive process based on real-time processing feedback, aiming to enable the output prediction value of the performance degradation assessment network to converge to the actual chemical state of the oil more quickly.

[0045] See Figure 5This is a deviation pattern recognition diagram for a hydraulic oil gradient centrifugal purification system, used to quantitatively analyze the dynamic changes in the probability of deviation sources over time throughout the purification process. From 0 to 60 minutes, the probability of insufficient centrifugal force intensity first decreases and then increases, reaching a peak around 45 minutes, becoming the dominant deviation source in this stage. The probability of improper grading sequence is approximately 0 throughout the process, indicating that the initial process sequence design is reasonable and there are no obvious stage switching issues. The probability of inaccurate identification of degraded components remains at a low level of 0.1-0.15, as this stage mainly focuses on physical impurity removal, and degraded component separation has not yet started, so the identification error has minimal impact. From 60 to 120 minutes, the probability of improper grading processing sequence rises rapidly, reaching a peak around 75 minutes, becoming the dominant deviation source in this stage. The probability of insufficient centrifugal force intensity drops rapidly to below 0.1, indicating that the centrifugal force has reached the target intensity and is no longer the main deviation source. The probability of inaccurate identification of degraded components remains approximately 0, indicating that degraded component separation has not yet become a core step. The probability of inaccurate identification of performance-degraded components in the 120-180 min period spikes to 0.65 at the 125 min mark, then slowly decreases, becoming the sole dominant source of bias in this stage. The other two probabilities drop to approximately 0, indicating that the physical impurity removal process has been completed and no longer generates bias.

[0046] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A method for on-line purification and performance regeneration of hydraulic oil regulated by centrifugal force field, characterized in that, include: Real-time operating parameters of the hydraulic system and real-time status data of the hydraulic oil are collected online. Based on the real-time operating parameters and real-time status data, the concentration distribution of different contaminants and the content of performance-degrading components in the hydraulic oil are predicted to obtain the oil condition assessment results. Based on the oil condition assessment results, the intensity distribution curve of the required centrifugal force field is dynamically calculated using a state inversion algorithm. The intensity distribution curve is used to guide the generation of gradient-changing centrifugal separation force. Based on the intensity distribution curve, the centrifuge device is controlled to generate a suitable directional centrifugal force field, which provides differentiated separation driving force according to the physical characteristics of the pollutants; Under the action of the directional centrifugal force field, the hydraulic oil is subjected to gradient centrifugal classification and purification treatment, and solid particles, free water and colloidal substances are removed step by step by adjusting the centrifugal force step by step. During the simultaneous process of gradient centrifugal grading for oil purification, the local intensity of the centrifugal force field is finely adjusted based on the content of performance-degraded components identified by the state inversion algorithm, in order to selectively remove degraded oil molecules while retaining intact base oil and effective additives.

2. The method for online oil purification and performance regeneration of hydraulic oil controlled by centrifugal force field as described in claim 1, characterized in that, Based on the real-time operating parameters and the real-time status data, the concentration distribution of different contaminants and the content of performance-degrading components in the hydraulic oil are predicted, including: The real-time operating parameters include system pressure and oil temperature, and the real-time status data includes contamination level data, viscosity data, and acid value data. The real-time operating condition parameters and the real-time status data are input into a pre-trained pollutant association inference network, which is a computational model containing multiple nonlinear transformations. The pollutant association inference network maps the predicted concentration of solid particulate matter, the predicted content of emulsified water, and the predicted content of colloidal substances in different particle size ranges. Simultaneously, the viscosity data and acid value data from the real-time status data are input into the performance degradation assessment network, which is used to correlate changes in the chemical structure of the oil. The content of oxidized polymers, acidic substances, and additive loss products in the oil is calculated using the performance degradation assessment network, which is the content of the performance degradation components. The predicted concentration values ​​are combined with the content of the performance-degrading components to form a complete oil condition assessment result.

3. A method of on-line oil purification and performance regeneration of hydraulic oil regulated by centrifugal field as claimed in claim 2, characterized in that, Based on the oil condition assessment results, the intensity distribution curve of the required centrifugal force field is dynamically calculated using a state inversion algorithm, including: The predicted values ​​of pollutant concentrations and the content of performance-degrading components in the oil condition assessment results are used as the input vectors for the condition inversion algorithm. Establish a set of inversion equations describing the relationship between centrifugal force field intensity, interaction time, and removal efficiency of specific pollutants; Substitute the input vector into the inversion equation set and solve in reverse the process to obtain the centrifugal force field intensity required to achieve the target removal efficiency as a function of time or spatial position; The obtained function is smoothed and piecewise fitted to generate the intensity distribution curve that changes continuously within a preset processing period. The intensity distribution curve clarifies the intensity value that the centrifugal force field should reach in different processing stages.

4. The hydraulic oil on-line purification and performance regeneration method regulated by centrifugal field as claimed in claim 3, characterized in that, Based on the intensity distribution curve, controlling the centrifuge device to generate a suitable directional centrifugal force field includes: The intensity distribution curve is converted into a sequence of spindle speed control commands for the centrifugal device. According to the spindle speed control command sequence, the speed of the drive motor is adjusted so that the centrifugal drum produces a speed change that matches the intensity distribution curve; Inside the centrifugal drum, based on the characteristics of the difference in centrifugal force at different radii, and combined with the density and particle size information of the pollutants, the optimal separation zone for different pollutants is planned. By adjusting the spatial position and flow field distribution of the oil inlet, the hydraulic oil to be treated is guided to the corresponding optimal separation area, thereby forming a directional centrifugal force field inside the centrifugal drum that matches the contamination characteristics.

5. A method of on-line oil purification and performance regeneration of hydraulic oil regulated by centrifugal field as claimed in claim 4, characterized in that, Under the action of the directional centrifugal force field, the hydraulic oil undergoes gradient centrifugal classification and purification treatment, including: In the initial stage of gradient centrifugal classification oil purification, a relatively low initial centrifugal force field intensity is used to allow the densest and largest solid particles in the hydraulic oil to settle and separate first in the directional centrifugal force field. After the initial stage is completed, the centrifugal force field intensity is increased to the intermediate intensity range according to the intensity distribution curve; Under the aforementioned medium strength range, fine particles with low density, demulsified free water, and some macromolecular colloidal substances are removed from the hydraulic oil; In the final stage of gradient centrifugal grading oil purification, the intensity of the centrifugal force field is adjusted to the highest intensity within a preset range according to the intensity distribution curve. At the highest intensity, residual fine particles, emulsified water, and soluble colloids are further removed to complete the graded oil purification process.

6. The method for online oil purification and performance regeneration of hydraulic oil controlled by centrifugal force field as described in claim 5, characterized in that, During the synchronous process of gradient centrifugation and grading for oil purification, the local intensity of the centrifugal force field is fine-tuned based on the content of performance-degrading components identified through a state inversion algorithm, including: In the final stage of the gradient centrifugal classification oil purification process, a simultaneous separation procedure for performance-degraded components is initiated. Based on the content of performance-deteriorating components in the oil condition assessment results, calculate the additional centrifugal force increment required for selective separation; Based on the highest intensity specified by the intensity distribution curve, the additional centrifugal force increment is superimposed to form a reinforced separation force field targeting the performance-deteriorating component; Under the enhanced separation force field, oxidized and polymerized macromolecules, some acidic products, and ineffective additives are enriched and separated during centrifugation, while intact base oil and effective additives are retained in the purified oil.

7. The method for online oil purification and performance regeneration of hydraulic oil controlled by centrifugal force field as described in claim 6, characterized in that, Also includes: Real-time monitoring of the cleanliness and key performance indicators of hydraulic oil after gradient centrifugal grading purification, generating feedback data of the purification process; The feedback data from the oil purification process and the oil condition assessment results are input into the collaborative control logic model, which is used to optimize the centrifugal force field control strategy. The intensity distribution curve and the operating parameters of the gradient centrifugal grading oil purification process are adjusted in a closed loop using the aforementioned collaborative control logic model. The process is iteratively executed until the cleanliness and key performance indicators of the hydraulic oil reach the preset standards, thus completing the online oil purification and performance regeneration treatment of the hydraulic oil. The real-time monitoring of the cleanliness and key performance indicators of the hydraulic oil after gradient centrifugal grading purification includes: An online particle counter and a moisture sensor are installed in the clean oil outlet pipeline of the centrifuge to continuously acquire the number of particles and the water content of the treated hydraulic oil as the cleanliness index. A portion of oil sample is drawn out from the upstream branch of the clean oil outlet pipeline and flows through an online viscometer and acid value sensor to measure the kinematic viscosity and total acid value of the treated hydraulic oil, which are used as the key performance indicators. The data streams of particle number, water content, kinematic viscosity and total acid value are timestamped and packaged to generate the oil purification process feedback data.

8. The method for online oil purification and performance regeneration of hydraulic oil controlled by centrifugal force field as described in claim 7, characterized in that, The feedback data from the oil purification process and the oil condition assessment results are input together into the collaborative control logic model, including: An error assessment mechanism is established that uses the feedback data from the oil purification process as real-time observations and the oil condition assessment results as prediction benchmarks. By comparing the real-time observed values ​​with the predicted baseline using the collaborative control logic model, the deviation between the current oil purification effect and the expected target is calculated. Based on the calculated deviation, determine whether the deviation is due to insufficient centrifugal force field strength, improper grading sequence, or inaccurate identification of performance-degrading components. Based on the judgment of the source of deviation, the collaborative control logic model generates a set of control instructions including the adjustment amount of centrifugal force field intensity, the adjustment amount of graded stage duration, or the correction amount of performance degradation component identification parameters.

9. The method for online oil purification and performance regeneration of hydraulic oil controlled by centrifugal force field as described in claim 8, characterized in that, The collaborative control logic model is used to adjust the intensity distribution curve and the operating parameters of the gradient centrifugal grading oil purification process in a closed loop, including: The centrifugal force field intensity adjustment amount in the control instruction set is fed back to the input parameters of the state inversion algorithm, triggering the state inversion algorithm to recalculate the new intensity distribution curve; The duration adjustment of the grading stage is directly applied to the timing controller of the gradient centrifugal grading oil purification process to adjust the duration of each stage of the process. The correction amount of the performance degradation component identification parameter is input into the internal weight matrix of the inference calculation model to fine-tune the performance degradation assessment network online, so as to improve the prediction accuracy of the content of performance degradation components. Based on the newly generated intensity distribution curve, the adjusted treatment duration, and the fine-tuned performance degradation assessment network, a new round of online hydraulic oil purification and performance regeneration treatment cycle is initiated.

10. The method for online oil purification and performance regeneration of hydraulic oil controlled by centrifugal force field as described in claim 9, characterized in that, The steps for constructing a collaborative regulation logic model include: Construct a closed-loop control network architecture with the feedback data of the oil purification process as the observation input, the oil condition assessment results as the reference input, and the control commands as the output; In the closed-loop control network architecture, a state comparator is designed to compare the real-time observed values ​​of cleanliness indicators and key performance indicators in the oil purification process feedback data with the corresponding predicted values ​​of pollutant concentration and performance degradation component content in the oil condition assessment results, and calculate the deviation vector between each predicted value and the real-time observed value. Based on the aforementioned deviation vector, the weighted influence coefficients of each deviation on the oil purification effect are calculated using a fuzzy inference engine. Using a multi-objective decision-making algorithm, the weighted influence coefficients, preset energy consumption constraints, and processing efficiency constraints are comprehensively calculated to generate a candidate instruction set that includes the centrifugal force field intensity adjustment, the graded stage duration adjustment, and the performance degradation component identification parameter correction. The strategy evaluation module simulates and evaluates the execution effect of each control instruction in the candidate instruction set, selects the optimal control instruction combination that makes the predicted processing result of the next processing cycle approach the preset standard the fastest, and outputs the optimal control instruction combination as the control instruction set of the collaborative control logic model.

Images